Of the 152 dwarf F, plants grown at Bozeman in 1923, 114, or 75 per cent, matured seed and 38, or 25 per cent, did not head and produce seed and were classed as grass clumps. Of the first class 25 were grown at Bozeman in 1924. The breeding behavior of these strains in the F, generation is shown in Table 5.

3

Of the 25 F, dwarf strains 4, or 16 per cent, bred true for dwarfness, and the remainder segregated in a ratio not significantly different from 3 dwarfs to 1 normal. Assuming that 16 per cent of the 114 dwarfs in the F, population would have proved homozygous if tested, it appears that there were about 18 fertile homozygotes to be added to the 38 sterile grass clumps, making 56 in all. Correcting the F, Bozeman results on a percentage basis of F, plants grown at that station, the figures shown in Table 6 are obtained.

TABLE 6.-Correction of the Bozeman F2 dwarf data on the basis of F3 results

This corrected deviation of 5±3.92 is a close fit to the expected 2:1 ratio, as it is within three times the probable error. It is important to note that at Bozeman in 1924 the four true-breeding F2 dwarf families all headed and matured seed. From these data it seems that the grass clumps appear only as a homozygous condition of dwarfness, the ability of the plants to head and mature seed being due to environmental conditions.

An interpretation of the above data can be satisfactorily made on a genetic basis by assuming DD a factor for dwarfs and NN an inhibitor or factor for normals. As Waldron (20) has shown, Kota Marquis hybrids produce dwarfs in ratios involving two or three genetic factors, and Clark (5) observed no dwarfs in Kota × Hard Federation crosses, the formula for Marquis could be DDNN and Hard Federation ddnn, or vice versa. A dominant third factor necessary for dwarfness may be present in Marquis and Hard Federation but lacking in Kota. As this would be common to the varieties dealt with here, it need not be represented in the formulas. The F, of the Marquis x Hard Federation cross will thus be represented as DdNn. The following formulas may represent the F2 genotypes and their breeding behavior in the F, generation:

1 DDNN

2 DdNN.
1 dd NN.

2 dd Nn

1 ddnn...

---

7.

4 Dd Nn..

2 DDNn.

2 Ddnn..

1 DDnn..

79182-26-2

Normals, breeding true to normals.

Normals breeding 13 normals to 3 dwarfs.
Normals, breeding 3 normals to 1 dwarf.
Dwarfs, breeding 3 dwarfs to 1 normal.
Dwarfs, breeding true to dwarfs.

The corrected numbers corresponding to this 7:4:2:2:1 genotypic ratio from the Bozeman F2 data of 752 plants are 340:172: 88:96:56. The calculated on the above ratio would be 329:188: 94:94:47. The deviations in each class are 11, 16, 6, 2, 9, from which the value of 0.42 for P indicates a very close fit.

AWNEDNESS

Biffen (2) was first to point out that the awnless condition of a wheat spike is dominant. Other early workers obtained similar results in the first generation and also found that in the second generation awnless and awned plants occur in the single Mendelian ratio of 31 or of 1:2:1 when intermediates occur. Howard and Howard (11) were the first to record two factors in the inheritance of awns when they grouped all awned and awn-tipped classes as awned, which in comparison with awnless gave a 151 ratio. They thus concluded the awned condition to be dominant.

2

Clark (5) has shown that in Kota-Hard Federation crosses two genetic factors did not entirely account for the breeding behavior in F, and F, after the material was classified into five classes; that is, (1) awnless, as Hard Federation; (2) apically awnletted; (3) awnletted; (4) short awned; and (5) awned, as Kota. The F, represented class 2, apically awnletted, and in F, all F, classes segregated into at least four classes. Two genetic factors would explain the behavior of strains breeding true for classes 1 and 2 and also those for classes 4 and 5, but an additional factor or factors would be necessary to account for the variation remaining.

2

2

In the present study (pl. 2), in which Hard Federation is awnless (class 1+), Marquis is awnletted (class 3-), and F, apically awnletted (class 2), the F, and F, material is separable into three classes, (1) awnless, (2) apically awnletted, and (3) awnletted. Class 1, awnless, usually is without awnlets in the apical portion of the spike, although, like Hard Federation (1+), it sometimes has a few awnlets 1 to 2 millimeters long. Class 2, apically awnletted, has awnlets 2 to 20 millimeters long at the apex of the spike, but rarely extending to the central and basal portions. Class 3, awnletted, has awnlets from 3 to 40 millimeters long, the shorter occurring at the base of the spike and the length increasing toward the apex, where many exceed the length of the awnlets on Marquis (3-) by 10 to 20 millimeters. It is seen from Plate 2, A and B, that selections of the hybrid have a greater awn development than Marquis.

The data by classes, obtained in F2 at Bozeman and Moccasin in 1923, are given in Table 7. Only normal plants are included in this study.

The data given in Table 7 show segregation into the three awnedness classes for 12 F, families, of which 6 were grown at Bozeman and 6 at Moccasin and of which 3 represent reciprocal crosses at each station. The results from reciprocal crosses at the two stations do not show significant differences. At Bozeman the Hard Federation X Marquis families in which the awnless Hard Federation parent was the female, have a slightly higher percentage of awnless plants, and the Marquis Hard Federation families in which the awnfetted Marquis parent was the female, have a slightly larger percentage of awnletted plants. At Moccasin the same relationship existed for the

awnless segregates, but there were much larger percentages of awnletted plants in the Hard Federation X Marquis crosses than in the reciprocal. There is no consistent or significant evidence, therefore, of maternal influence affecting the awnedness factors in reciprocal crosses of these wheats.

TABLE 7-Segregation in the F2 generation into three classes for awnedness of 1,050 plants of 12 Marquis Hard Federation wheats and reciprocal crosses grown at Bozeman and Moccasin, Mont., in 1923

Combining classes 1 and 2 and comparing with class 3, the data are shown to be close to a 3:1 ratio except for two F, families at Bozeman. By combining classes 2 and 3 in comparison with 1, the data from 12 families at two stations are close to a 15:1 ratio. These data indicate the presence of one and two genetic factor differences.

The segregation of the Bozeman material into the three classes conformed closely to a 1:12:3 ratio, and that of the Moccasin material was near a 1:11:4 ratio. It was therefore necessary to make a careful study of the F, material to determine the breeding behavior of the different classes. This study was made at Bozeman in 1924 on material grown from F, Bozeman seed. The behavior of the three classes in the F, generation is presented in Table 8.

TABLE 8.-Breeding behavior in the F3 generation of 12,197 plants from MarquisHard Federation wheat crosses of the three awnedness classes grown at Bozeman, Mont., in 1924

To interpret the data on a one-factor basis there must necessarily be approximately 25 per cent of true-breeding recessives. These recessives may be among those plants having either more or less awn development than the F1, which was apically awnletted (class 2) and intermediate between the two parents.

In the F, generation, combining classes 1 and 2, the plants as classified gave the percentages shown in Table 9.

TABLE 9.-Percentage totals of F2 results by combining classes 1 and 2 in comparison

Bozeman.
Moccasin.

Percentage of total.

with class 3

The awnletted class 3 from Bozeman material bred true in the F3 generation, and as shown in Table 8 nine families also bred true for class 3 that had been described in F, as belonging to class 2.